One of the early customers for Power*Star is planning to use it with a single-board computer (SBC) at the telescope, and asked me to customize the enclosure to allow mounting the SBC on top of P*S. It is more than worth noting here that where to locate the host computer is an important question. Some competitive products to P*S have a built-in processor (ranging from small Raspberry Pi boards to full-blown Windows machines with high-end Intel processors), while others take the same approach as P*S, providing only the features specific to astrophotography. While it is sometimes convenient to have everything integrated into a single unit, I decided not to include a processor in P*S because computers are commodity products. It is very easy to choose and purchase a laptop or other computer that is exactly what you need and far less expensive than what it would cost to design it into the P*S product. Furthermore, I would probably have to produce at least two different versions, since the processing needs of users varies tremendously. P*S doesn’t know or care where the host computer if physically located. A potential advantage in having it mounted at the telescope is that you can connect to the computer via Wi-Fi and avoid having a USB cable going from the telescope to the base of the mount (thus needing only a power cable). If the host computer is ‘off telescope’ it can be some distance away (as far as you can reach with a USB cable), so it could be a laptop that you use alone to control and monitor the image capture process, or it can be a ‘headless’ computer that is connected via Wi-Fi or wired ethernet to a separate computer, using remote desktop software.
In this case the SBC is a LattePanda V1, which is based on an Intel CherryTrail processor, a multi-core, modern version of the old Atom processor. I used to run all my image capture software on a ‘netbook’ with the original, single-core Atom processor, and it worked flawlessly, so I have no doubt that the LattePanda (LP) can handle the task for ALMOST all users. One of the possible exceptions is me. Specifically, I use a Paramount MyT equatorial mount, which requires the use of their software (called TheSkyX, or TSX), which is a bit of a processor hog. I’ve found a very compact Core i3 computer from Lenovo that runs it adequately, but it’s rather heavy to mount at the telescope, so I have placed it at the base of the mount.
It turned out that the best way to attach the LP to P*S required some modification to both enclosures. First, I added 4 ‘stand-offs’ to the top of the P*S enclosure:
These thread directly into the top surface using M3 threads.
The LP enclosure has ventilation holes in the bottom, so I had to leave some space for air flow. But it needed to be raised up a bit anyway because the LP enclosure doesn’t quite fit between the front and back covers of the P*S enclosure. The LP enclosure is usually held together with screws that are accessed through the bottom, but I was able to modify it (pretty easily, actually) to open from the top instead, which makes it much more convenient.
Above we see the bottom half of the LP enclosure attached to the stand-offs with M3 machine screws. It probably would have been better to use flat head screws, but this works and is a bit easier to do.
The LP board is then placed in its enclosure and secured with 4 more stand-offs that thread into the bottom half of the enclosure. The stand-offs are up-side-down relative to the original design. By using a slightly larger thread size for the stand-offs, I was able to tap threads into what had been clearance holes for screws coming up from the bottom. What had been threaded holes in the top cover were then drilled out to be clearance holes for screws coming down from the top (see first photo). In this case I did use flat-head screws to maintain a flat surface over the top of the enclosure.
A very similar customization could be used to attach a Raspberry Pi enclosure to the top of P*S. Larger units, such as an Intel NUC, or the Lenovo computer I am using would be a bit more difficult due to the greater weight and complexity of such enclosures. Powering the computer is also an important consideration: Both the LP and Raspberry Pi SBCs need 5V power, while both the NUC and Lenovo computers require non-standard DC voltages (around 20V). So while the LP and R-Pi can be powered using the variable voltage output of P*S (set to 5V), the others would require either a separate DC-DC converter (going from 12V to ~20V), or running a separate cable to an AC adapter at the base of the mount (it’s generally not a good idea to run AC power up to the telescope).
Customizations like this are a major advantage to being a very small manufacturer. I am able to do this for my customers, and relatively inexpensively. Mass production generally makes it unthinkable to do this kind of customization. There are other customizations possible with P*S as well. For example, if you have a power source that makes it essentially impossible to connect power with the wrong polarity, it can be advantageous to bypass the reverse polarity protection in P*S because this circuit is the 2nd largest cause of voltage drop in a typical installation (the first being the power cable, even with the beefy 14 gauge cable I provide with P*S).
If you have other ideas or problems to solve to use P*S, email me with a description. I’ll do my best to make it work!